STL Video Transcripts

This repo is for people to contribute to editing transcripts of the STL Open Code Review meetings. Thank you!

Adding Captions

To add captions, first make sure to pipe up in the correct issue thread, and inform us that you're taking a video on, so that we don't duplicate work. Then, you can start on a pull request.

You should start with metadata.json; add a "name" and a "description" in accordance with that section.

Then, copy original.vtt to edited.vtt, and you can start editing there in accordance with Captioning Style. You can get the video at the link in metadata.json ("private-link"). If you need any assistance at all, don't hesitate to ask either on the STL discord, or on the issue thread.

Layout of the Repo

Each video transcript is in its own directory: transcripts/<yyyy>/<yyyy-mm-dd>. There are three files in here:

• metadata.json: contains information about the video, including video title, description, and links
• original.vtt: the original machine-generated transcript. Do not edit this.
• edited.vtt: the edited transcript; this is the one you modify.

metadata.json

There are currently four fields in metadata.json:

• "name": Should be GH-NNNN (with NNNN being replaced by the number), followed by a comma, followed by the actual title of the PR. If this is a part two, put (part 2) in a parenthetical after the PR number, but before the comma.
  • "GH-2158 (part 2), Extend memcpy, memmove, and memcmp optimizations"
  • "GH-2019, <functional>: Implement invoke_r"
• "description": Should look something like:
  • "In this video we review GH-2158: https://github.com/microsoft/STL/pull/2158, titled \"Extend memcpy, memmove and memcmp optimizations\", now on part 2. Part 1 of the code review can be found at https://youtu.be/aQ4isF4f0bY.",
• "private-link": Will be filled out by maintainers; the link to the unlisted video
• "public-link": Will be filled out by maintainers; the link to the listed video on the Microsoft official youtube channel

Captioning Style

The following rules should be used:

• Remove any NOTE Confidence lines
• To denote a period of silence while mouths are moving, (for example, if someone forgot to unmute their microphone), or for noises or notes, use [] brackets:

  00:00:03.000 --> 00:00:08.290
  [silence]

  01:14:59.310 --> 01:15:02.340
  [under breath] thought it was here, _Can_strong_order.

• If what was said was just completely unintelligible, write [unintelligible]:
• For greater than, less than, and ampersand symbols, use the &blah; syntax:

  00:00:49.370 --> 00:00:52.286
  line, right line 4005 of <xutility>.

• To denote a new speaker, use two greater than symbols, followed by the given name of the person speaking with the first letter capitalized:

  00:00:10.940 --> 00:00:12.896
  >> Stephan: Cool, hi, welcome back to another

• For spoken references to code, write in the captions as you would write code; for example, if someone says "this takes a const ty one ref left and a const ty two ref right", write:

  01:11:11.022 --> 01:11:14.031
  This takes a const _Ty1& _Left and a const _Ty2& _Right.

• For unicode characters, enter them as utf-8; for example (that is an ellipsis):

  01:12:48.350 --> 01:12:50.840
  On line 3…, no, line 700.

• Use an ellipsis instead of three periods
• Instead of using a hyphen-minus or an em dash, use an en dash to separate thoughts or for parentheticals. However, for hyphenated words, use a hyphen-minus:

  01:13:44.380 --> 01:13:46.200
  Argument Dependent – and I don't do this for everything,
  
  01:13:46.200 --> 01:13:47.864
  like SFINAE, but may as well.
  
  01:13:47.864 --> 01:13:50.400
  This is also not universally known – Lookup.

For a complete example, take a look at 2021-09-09.